Genre Variation and Changes in Frame Sequences Across Cultures: The Case of Criminology RA Abstracts in English and French

Abstract Though not as widely studied as the Research Article (RA), the abstract has attracted increasing interest among researchers over last decades (Swales 1990, Bhatia 1993, Dos Santos 1996, Lorés-Sanz 2008, Bondi/Cavalieri 2012, Cavalieri 2014). A number of contrastive or comparative studies of abstracts in English and other languages (Martín-Martín 2005, Lorés Sanz 2006, Van Bonn & Swales 2007, Diani 2014) have already been carried out considering mainly the hard sciences and some soft sciences such as linguistics and history, however no cross-cultural analyses have been conducted so far between RA abstracts in English and RA abstracts in French published in the legal field. This paper seeks to investigate genre variation and changes in frame sequences comparatively in English and French RA abstracts from criminology journals. Using a genre analytical approach to qualitative and quantitative data, the paper reports on two comparable corpora, i.e. English and French, of electronically retrieved abstracts from Criminology Journals published in 2014. The two corpora are composed of three journals per language, namely Criminology, Journal of Criminal Justice, Journal of Criminal Law and Criminology for the English corpus, and Champ Pénal, Criminologie, Revue Canadienne de Droit Pénal et Criminologie for the French corpus. The analysis will be carried out following two main steps, i.e. a macro-analysis and a micro-analysis. In the former step, the corpora are compared by the analysis and discussion of the basic IMRD rhetorical move structure for the RA often proposed in the literature (e.g. Nwogu 1990; Swales 1990; Bhatia 1993; Ventola 1994; Martín-Martín 2002) and the additional five moves model postulated by Dos Santos (1996) with the aim of investigating the linguistic and rhetorical variation in the abstract genre from a cross-cultural perspective. In the latter, we look at frame sequences (Bondi/Cavalieri 2012) combining forms of self-mentions and frame markers (Hyland 2005), i.e. personal patterns (e.g. we argue / nous questionnons), impersonal patterns (e.g. it is argued / il est question) and locational patterns (e.g. the paper argues / l’article questionne) (Dahl 2004). Provisional results show that the abstracts under investigation largely follow the international conventions based on the norms established by the English-speaking international academic community. However, variation across the two cultures emerged from the linguistic realizations of framework sequences. Cross-cultural implications are discussed at the close.


Introduction
Academic discourse has been the centre of increasing scholarly attention in the last two decades, especially from a genre perspective (Swales 1990(Swales , 2004Bhatia 1993Bhatia , 2004. Though not as widely studied as the Research Article (RA), the abstract has attracted increasing interest among researchers investigating academic discourse in recent times (Swales 1990, Bhatia 1993, Kaplan et al. 1994, Dos Santos 1996, Hyland 2000, Bondi 2001, Stotesbury 2003, Pho 2008, Lorés-Sanz 2004, 2006, Dahl 2009, Golebiowski 2009, Gillaerts/Van De Velde 2010, Bondi/Cavalieri 2012, Cavalieri 2014. As Bondi (1997: 396) states, "abstracts would seem to provide excellent material for genre analysis. Their textual structure is comparatively easy to identify and their size is manageable for different types of linguistic analysis. The linguistic literature on abstracts, however, is not so extensive as one would expect and most contributions tend to focus on textual structure only".
A number of contrastive or comparative studies of abstracts in English and other languages (Martín-Martín 2002and Lorés-Sanz 2009 for Spanish, Diani 2014 for Italian, Busch-Lauer 1995 for German, among others) have already been carried out considering mainly the hard sciences and some soft sciences such as linguistics and history, however no cross-cultural analyses have been conducted so far between RA abstracts in English and RA abstracts in French published in the legal field (law abstracts have only been compared to business ones in English (Hatzitheodorou 2014)). As a matter of fact, to the best of our knowledge, we found only comparative analyses between English and French abstracts of IT (Crosnier 1993) and linguistics (Van Bonn/Swales 2007) research articles, or concerning other academic genres such as the research article itself (Dahl 2004) or PhD thesis abstracts (Bordet 2011, Wable/Holzem 2004.
Therefore, the present study seeks to investigate genre variation and changes in framework sequences as well as in rhetorical structure comparatively in English and French RA abstracts from criminology journals. We aim at exploring cross-cultural changes in both communicative practices and linguistic patterns, in order to observe whether French abstracts show some traces of the influence of the norms of the English academic discourse community on the genre. Differences in writers' self-awareness in the two academic cultures will be also considered.
Interesting in the study of legal discourse is the research carried out by Goźdź-Roszkowski (2011) into seven genres of legal texts in American English: (i) academic journals (71 texts amounting to a total of 552,487 words), (ii) briefs (64 texts amounting to a total of 763,222 words), (iii) contracts (177 texts amounting to a total of 1,178,616 words), (iv) legislation (60 texts amounting to a total of 1,178,516 words), (v) opinions (114 texts amounting to a total of 1,182,246 words), (vi) professional articles (100 texts amounting to a total of 201,404 words), and (vii) textbooks (104 texts amounting to a total of 519,116 words). The integration of corpus and discourse analysis tools proposed in that work is also taken as basis for the present study. Although the methodology used is worth to be noted, it seems not to suit our analysis since our work deals with a small corpus. Moreover, differently from Goźdź-Roszkowski's work, our focus is on the academic discourse used by law experts and not on the construction of professional discourse through different genres.
The paper is structured as follows: section 2 describes the comparable corpora, the methods and the frameworks of analysis followed. The basic moves (Bhatia 1993), the variations to the IMRD (Introduction, Methodology, Results, Discussion) model (Dos Santos 1996, Hyland 2000, the different types of abstracts (Swales/Feak 1994, Lorés-Sanz 2004, the use of metadiscourse (Hyland 2005) are some of our working notions. Section 3 turns to data analysis from two perspectives: on the one hand, we compare the generic structure of English and French criminology abstracts; on the other hand, we consider the use of frame markers in representing the structure of the abstracted article and the mentions to author's identity. Section 4 offers some concluding remarks on the comparative results.

Materials and methods
Using a genre analytical approach to qualitative and quantitative data, this study reports on two comparable corpora, i.e. English and French electronically retrieved abstracts from Criminology Journals published in 2014. The two corpora are composed of three journals per language, namely Criminology (C), Journal of Criminal Justice (JCJ), Journal of Criminal Law and Criminology (JCLC) for the English corpus, and Champ Pénal (CP), Criminologie (C), Revue française de Criminologie et Droit Pénal (RFCDP) for the French corpus. The choice of these three francophone journals has been determined by the fact that, although they require abstracts in French and English (Criminologie adds a Spanish version), the associated articles are written in French. This lets us suppose that the French version of the abstract is the original one, while English versions are translations. They can, in fact, be considered as author abstract, i.e. abstracts written by the same author of the related RA (Bordet 2011: 30, Busch-Lauer 2014. From a first quantitative overview of the two corpora, it is possible to notice that the number of abstracts retrieved for English is higher than the number of French ones (see table 1). As we have already said, we decided to collect the abstracts published in 2014 to have a coherent set of data for the comparative analysis. However, this resulted in a big difference in the number of abstracts for each corpus, i.e. 85 vs 36 for the English and for the French corpus respectively. As a consequence, to normalize the comparison of our linguistic data, the results of our subsequent analysis will be given in percentage. The wider presence of English abstracts is also confirmed by the difference in the size (number of tokens) of the two corpora: The reduced size of the French corpus is probably due to the dominant role of English in international scientific communication (Tardy 2004, van Weijen 2012) as it reaches a much wider audience (Diani 2014). In fact, roughly 80% of all the journals indexed in Scopus are published in English (van Weijen 2012: 7).
For this study, our methodology is based on the integration of corpus and discourse perspectives focusing on two main points: a) the structure of the abstract, in relation to the nature of the article; b) the use of frame markers in representing the structure of the abstracted article. The analysis was thus carried out following two main steps, i.e. a macro-analysis (using discourse analytical tools) and a micro-analysis (using corpus analytical tools [i.e. Antconc for the study of wordlists, concordances and collocations]), with the aim of investigating the linguistic and rhetorical variation in criminology abstracts from a cross-cultural perspective.
In the former step of our research, the corpora were compared focusing on the analysis and discussion of the basic IMRD rhetorical move structure for the RA. Following Bhatia (1993: 78-79), we started from the assumption that four basic moves can be identified in RA abstracts: 1. Introducing Purpose (= Introduction): a statement of the author's aims (hypothesis to work on, the purposes of the paper, the thesis supported); 2. Describing Methodology (= Methodology): a description of the experimental design (data, procedures, methods, scope); 3. Summarising Result (= Results): a statement of the author's findings and observations on the suggested solutions to the problems; 4. Presenting Conclusions (= Discussion): the interpretation of the results, a statement of the inferences that can be drawn from the findings and the implications and applications they suggest. Some variants to this move structure have been added later to account for a separate Background Move (BM) before "Introducing Purpose", called Situating research (Dos Santos 1996) or Introduction move (Hyland 2000), providing a justificatory context for the paper and a research motivation. Variation among types of abstracts has also been recognized. Lorés-Sanz (2004), for example, considers the traditional distinction between informative and indicative abstracts, that will be discussed later under the labels of argumentative and empirical abstract respectively. Informative/empirical abstracts are often characterized by a full IMRD structure following the organization of the scientific RAs they are associated with. Indicative/argumentative (or descriptive) abstracts only focus on the purpose of the RAs, without presenting results. These seems to correspond rather to the CARS model (Create A Research Space) introduced to study the structure of RAs Introductions (Swales 1990): 1. Establishing a Terri-tory; 2. Establishing a Niche; 3. Occupying the Niche. Patterns of the two models can sometimes be mixed. This theoretical frameworks allowed us to check if and to what extent the moves are present in both the English and the French corpora.
In the latter step, we investigated the rhetorical strategies and the use of metadiscursive phraseology (Hyland 2005) connected with the different moves, paying particular attention to framework sequences (Bondi 2010, Bondi/Cavalieri 2012 combining forms of self-mentions and frame markers, i.e. personal patterns, impersonal patterns and locational patterns (Dahl 2004). On this basis, we identified three main styles in abstracts: personal, impersonal and mixed. 1) Impersonal style: abstracts including impersonal (it is argued that / il est question de), locational (the paper sets out to, the article deals with / cet article porte sur, cette recherché tente de), locative (an attempt is made in this article to infer / dans cet article nous comparons) and 3 rd person patterns (the author suggests / l'auteur considčre); 2) Personal style: abstracts in which the authors systematically make use of the 1 st person singular and plural pronouns (I argue, we conclude, my analysis / on conçoit, nous proposons, notre démarche s'appuie); 3) Mixed style: abstracts using both impersonal and personal patterns.

Results
This section is divided into two parts: the first part will give results of the analysis of the generic structure of RA criminology abstracts in English and French, in relation to the nature of the article whether empirical research report or argumentative discussion of an issue; the second part will take into account writer's identity through the lens of the abstract style and will deal with metadiscursive expressions, mainly focusing on framing verbs.

Generic structure
Two basic types of abstracts can be observed: a) argumentative: the focus is on making a claim and defining an issue in the disciplinary context (examples 1a and 1b); b) empirical: the focus is on reporting on the research carried out (examples 2a and 2b). These two categories show strong similarities with the traditional distinction between informative and indicative abstracts defined by Lorés- Sanz (2004). 1a) ARGUMENTATIVE In a recent article published in Criminology, Burt and Simons (2014) claimed that the statistical violations of the classical twin design render heritability studies useless. Claiming quantitative genetics is "fatally flawed" and describing the results generated from these models as "preposterous," Burt and Simons took the unprecedented step to call for abandoning heritability studies and their constituent findings. We show that their call for an "end to heritability studies" was premature, misleading, and entirely without merit. Specifically, we trace the history of behavioral genetics and show that 1) the Burt and Simons critique dates back 40 years and has been subject to a broad array of empirical investigations, 2) the violation of assumptions in twin models does not invalidate their results, and 3) Burt and Simons created a distorted and highly misleading portrait of behavioral genetics and those who use quantitative genetic approaches. (

2a) EMPIRICAL
This study compared and contrasted the views of formal and informal crime control among college students from China, Japan, and the U.S., and examined the correlates behind the views. Using the same questionnaire, this study collected data from 1,275 completed surveys in the three nations. The study revealed that both Chinese and Japanese respondents evaluated formal and informal control and their combination in crime control as more important than American counterparts did. The variable trust in police was a predictor of attitudes toward formal control and the mix of formal and informal control in all the three nations. Demographics in the U.S. were more important factors than in China and Japan in predicting the respondents' ranking of the importance of formal control and informal control and their combination in crime control. This is the first empirically comparative study of the perceived importance of formal and informal mechanisms in crime prevention and control in China, Japan and the U.S. The study found both similarities and differences in the perceived importance and reasons behind them. More research is needed in the future. (Journal of Criminal Justice) 2b) EMPIRICAL On observe dans plusieurs pays une forte prévalence des victimisations interpersonnelles chez les enfants et les adolescents. En Espagne, surtout parmi les groupesà risque, lesétudes n'ont pas réussià obtenir des profils complets de victimisation dans leséchantillons provenant de la communauté. Dans ce contexte, l'objectif de l'étude est de présenter des statistiques sur la portée, la nature et les tendances de la victimisation chez des adolescents espagnols patients ambulatoires. L'échantillon est composé de 148 adolescents en soins psychologiques. Un largeéventail d'expériences de victimisation ontétéévaluées en utilisant le Juvenile Victimization Questionnaire. Un pourcentageélevé de participants ont rapporté une certaine forme de victimisation interpersonnelle au cours de l'année précédente (84,5%), la plus fréquenteétant celle dles délits communs (62,8%). Les filles présentent un pourcentage plusélevé de victimisation en ligne que les garçons (22,1% et 7,5%). De plus, la polyvictimisation aété observée chez une proportion importante d'adolescents (29%). L'identification des polyvictimes en milieu clinique est une nécessité sociale car ces enfants et adolescents ont besoin d'interventions adaptées afin d'aider a prévenir de nouvelles expériences de victimisation et le développement de problčmes psychologiques. (Criminologie) Table 3 shows a more balanced subdivision between the two types of abstracts in French (with a slight predominance of empirical ones), while English abstracts seem to follow more frequently a structure typical for the hard sciences. Traditionally, academic French has given greater emphasis on discussing theoretical and methodological issues but today, probably due to the influence of English academic writing conventions, the French corpus shows an interesting presence of a more empirical research perspective, in fact greater importance is given to results and discussion. A further step of the analysis of the textual structure of abstracts consists in observing the presence of the moves. Table 4 shows that empirical abstracts both in English and French feature the full IMRD structure, while argumentative abstracts sometimes break with the rule, with a slight increase as far as French is concerned.  Both English and French abstracts tend to foreground the basically argumentative nature of academic discourse in a substantial background section, irrespective to the type of abstract whether argumentative or empirical. This seems to justify and motivate the relevance of the paper to the peer audience, but also to display a sense of the academic "community", positioning the new research within its boundaries, as both an "addressee" and a "warrant" of the research. Moreover, due to the growing competitiveness among researchers, their claims need to be supported by previous pieces of research carried out in the field, that can also highlight the centrality of the topic dealt in the paper (cf. Bondi/Cavalieri 2012: 51).
Besides studying the presence of a Background Move, it is also interesting to look at the sequences in which moves appears in the abstracts. As stated in table 4 above, in the English corpus only 5 (argumentative) abstracts out of 85 do not show a full IMRD sequence. The empirical abstracts, on the other hand, respect the IMRD structure displaying a high degree of regularity and orderliness in the development of the moves and thus of the rhetorical construction of the genre. In order to analyse if French is actually influenced by English academic discourse, we built the following table 6, which allows highlighting whether the classical sequence BM + IMRD is respected. Note that in abstracts the acronym IMRD -coined for RAs -must be replaced by the acronyms of Bathia's (1993) moves: 1. Introducing Purpose (IP); 2. Describing Methodology (DM); 3. Summarising Results (SR); 4. Presenting Conclusions (PC).
As we can notice, only 2 empirical abstracts out of 8 in Revue française de Criminologie et Droit Pénal and only 2 empirical and 1 argumentative abstracts out of 5 in Champ Pénal show a full IMRD structure. On the other hand, Criminologie seems to be more regular in following a standardised move structure: only 2 argumentative abstracts out of 23 depart from the rule.
The Background Move appears in both empirical and argumentative abstracts from the three journals, but unsystematically and, except for Criminologie (where only two abstracts are not regular) the moves do not follow the classical sequence. Finally, if it is often true that argumentative abstracts Table 6 Move structure in French abstracts  tend to focus on the purpose or seldom on methodology and/or conclusion, without presenting results, we must acknowledge that 2 of them surprisingly focus only on results.

Writer's identity
The use of metadiscourse (Hyland 2005) is connected with the different function of the moves. Therefore we paid attention to metadiscursive phraseology, in particular to framework sequences (Bondi 2010, Bondi/Cavalieri 2012 combining forms of self-mentions and frame markers (Hyland 2005) as in the following examples from the two languages under investigation: "the paper examines the implications" / "cet article propose un repérage"; "we suggest an integration" / "nous proposons une definition".
As we can observe in table 7, the presence of the writer's discursive identity in the text comes out of different forms of self-reference: nouns and/or pronouns referring to the paper or to the author, and some impersonal forms. These give raise to personal framework sequences or to impersonal framework sequences where the latter can be divided into 4 subcategories (Dahl 2004), (table 8).  It is now possible to consider lexical strings, moving on to collocations, patterns and semantic sequences (Hunston/Francis 2000, Hunston 2008), i.e. phraseology, integrating meaning, form and function. This section of the analysis was supported by Antconc 3.4.3 in order to investigate frequency and use of potentially metadiscursive expressions, pointing to interesting meta-discursive "nodes" of the abstract (wordlists and concordances).
The study deals in particular with the most frequent verbs of saying, divided into verbs of claim-making (English: suggest, argue, indicate, explain, find, infer, propose, show, support; French: proposer, montrer, analyser, synthétiser, affirmer, soutenir, définir) and verbs of topic-setting (English: examine, discuss, present, analyze, concern, consider, focus, investigate, report, review; French: utiliser, explorer, s'appuyer, s'intéresser, comparer, porter sur, presenter). Table 10 shows that French authors make a balanced use of the two categories of verbs saying. Table 10 Claim making vs. Toping setting in the two languages As far as the variation in their use, we can observe a mixed combination between personal and impersonal framework sequences and the two kinds of verbs of saying (tables 11, 12). For what concerns both claim making and topic setting patterns, it is possible to observe that both languages tend to prefer impersonal patterns, even though the difference between personal and impersonal sequences is more evident in French than in English where we find a more balanced situation. Moreover, English abstracts present a higher frequency of locational patterns, whereas the French ones of impersonals for claim making patterns and locational for topic setting patterns. This tendency of using impersonal structures may be considered as a sort of hedging. In fact, in this way, the authors may avoid direct responsibility for their claims by placing it on their papers (locational) or on a general impersonal pronoun (i.e. on, it).

Concluding remarks
The present study investigated genre variation and changes in framework sequences as well as in rhetorical structure comparatively in English and French RA abstracts from criminology journals. Our aim was to identify cross-cultural changes in both communicative practices and linguistic patterns, in order to observe whether French abstracts show some traces of the influence of the norms of the English academic discourse community on the genre. Moreover, we also dealt with the differences in writers' self-awareness in the two academic cultures.
A first preliminary observation of the two corpora showed a higher presence of empirical abstracts in both languages (with a more balanced situation in French) that corresponds to a more empirical perspective leaving aside the need to discuss theoretical issues. This is particularly significant for the French abstracts, if we consider that academic French has traditionally placed more emphasis on discussing theoretical and methodological issues.
This is also reflected in the investigation of the variation of genre structure and it is possible to say that the rhetorical structure of criminology RA abstracts in French conforms to the established norms of academic English when empirical abstracts are concerned. In fact, apart from some exceptions, they tend to feature a full IMRD structure. On the other hand, argumentative abstracts, which are a minority in English but also in French (even if the difference is less notable), are less standardized and tend to show a lower degree and regularity. Both English and French abstracts tend to foreground the nature of academic discourse in a background section. This signals the tendency to highlight the relevance of the paper indicating an increasing awareness of the importance of placing one's research within the debate that characterizes the discourse community. However, the presence of a Background Move does not seem to be related to the type, the style or the featuring of a full move structure. To sum up, we can say that the rhetorical structure presented by French abstracts is influenced by English academic conventions. In fact French abstracts show an interesting shift from the traditional theoretical focus to a research-based perspective giving greater importance to results and discussion sections.
As far as the analysis of framework sequences is concerned, the two corpora display a significant presence of locational patterns in both languages followed by impersonal self-mentions including passive forms, and personal patterns. Locational patterns allow researchers to place "in the middle" between personalization and impersonality. By using this strategy, they can talk about their papers rather than about themselves, but at the same time they can deal with their own research by retaining a more objective stance. A significant difference is the avoidance of the first person singular pronoun je, whereas English sometimes chooses the corresponding I. An equal distribution of claim making and topic setting framing verbs appears in French, whereas in English there is a higher frequency of topic setting verbal patterns. As for the latter, the French corpus presents a massive presence of locational patterns, whereas English is more personal.